If the beginnings of tone clusters are to be found in the adjacent use of semitones, consider two examples by Chopin (both the same, notes-wise and in their overall harmonic context); the first is the big chord on the final page of his B minor Scherzo, where not only does he give us a dominant seventh chord of C major (therefore including the notes G and F) over an F# pedal but then frantically reiterates that chord several times as if to drive home the dissonance, fortissimo - and the second appears just before the B major central section of his Polonaise-Fantaisie, where the same chord and the same F# underpinning (albeit with D, rather than G, as the uppermost note) gives a sense of disturbance before the tonality of B major is established.
OK, so then "that chord in Liszt's Via Crucis...
Of course, the difference in all these examples is that the clustering remains within some kind of tonal harmonic framework, whereas the much more extended us of tone clusters in the 20th century more often than not have quite different contexts, yet someting almost always grows out of something else, does it not?...
Best,
Alistair